3r9

An electron microscopical study of isolated mitochondrialmembranes treated with osmium tetroxide, potassium

permanganate, and formaldehyde

By K. DEUTSCH and W. KRAUSE

(Department of Electron Microscopy, University of Greifswald, East Germany)

and S. ROSENTHAL

(Department of Physiological Chemistry, Humboldt University, Berlin)

With 2 plates (figs, i and 2)

SummaryAn investigation of isolated mitochondrial membranes, fixed with osmium tetroxide,potassium permanganate, or formaldehyde, shows that a great number of smallparticles are attached to their surfaces. Sections of specimens fixed in potassiumpermanganate or osmium tetroxide appear to confirm these findings. An attempt hasbeen made to compare these particles with the 'elementary particles' described byFernandez-Moran and by Green, and also with particles revealed by negative stainingof mitochondrial membranes, as reported by Stoeckenius. Specimens fixed withformaldehyde, whether isolated mitochondria or material fixed in bulk, show verypoor contrast in electron micrographs.

IntroductionW E undertook the investigation of the effect of various methods of specimenpreparation on molecules of proteins, lipids, and nucleic acids, and also oncomparatively simple biological systems, such as isolated membranes,because we considered that these studies might contribute to the understand-ing of the effect of these methods on the cells themselves. We have alreadystudied the effect of certain fixatives on protein molecules (Deutsch, Fischer,and Krause, 1964), and we have now extended these investigations to isolatedmitochondrial membranes. We thought that a comparison of the effect ofvarious fixatives on these membranes might contribute to knowledge of theirstructure and of the chemical constitution of their structural elements. Wehave based our investigations on the information already available about thebiochemistry and ultrastructure of mitochrondria in sectioned cells, of isolatedmitochondria, and of isolated mitochondrial membranes (see especiallyWatson and Siekevitz (1956); Lehninger and others (1958); Lehninger andWadkins (1962)), and on the results obtained by Egger and Rapaport (1963)and by Rosenthal and Rapaport (1964) in their study of proteins andenzymes derived from the mitochondria of rat liver.

Materials and methodsThe membranes were obtained by the method devised by Schneider and

Hogeboom (1951). They were resuspended in a 0-25 M sucrose solution[Quart. J. micr. Sci., Vol. 105, pt. 3, pp. 319-323, 1964.]

320 Deutsch, Krause, and Rosenthal

(pH— 7-2), frozen at a temperature of —200 C, and stored at —50 C. Afterthawing they were homogenized ultrasonically (10 min, 800 kc/s, 5 watts/cm2).They were studied in the electron microscope, either without further treat-ment or after fixation. They were fixed with osmium tetroxide (2 min in a1% aqueous solution, buffered with sodium veronal/sodium acetate at pH 7),or with potassium permanganate (2 min in a 0-25% aqueous solution, bufferedwith sodium veronal/sodium acetate at pH 7), or with formaldehyde (2 min ina 5% aqueous solution of formalin, at pH 7). After fixation the membraneswere washed in distilled water. Small droplets of aqueous suspensions of themembranes (untreated, or treated with either osmium tetroxide, potassiumpermanganate, or formaldehyde) were deposited on specimen grids coveredwith collodion, and allowed to dry in air. The grids were studied eitherdirectly in the electron microscope, or after shadowing with platinum at anangle of 200. In another series of experiments, membranes that had beenfixed and washed as described above were dehydrated in ethanol, embeddedin a methacrylate mixture according to standard procedure, and sectioned atabout 30 mju, on an Ardenne-Westmeyer ultra-microtome. The electronmicrographs were taken with a Zeiss Elmi D 2 electron microscope at a voltageof 48 kV.

ResultsWe would like to emphasize that we are not only presenting micrographs

obtained with the 'good' fixatives, but also others, in order to demonstratethe difference between various methods, and because results obtained with'poor' fixatives may be used to confirm or even invalidate results obtained with'good' fixatives.

Figs. 1, D to F, and 2, A to E show patches of broken membranes. In someparts the electron density is much higher than in others, indicating thatseveral layers are lying on top of each other. Untreated specimens (fig. 1, D)and specimens treated with osmium tetroxide (fig. 1, F) or potassium per-manganate (fig. 2, A) do not show any special features; the contrast of speci-mens fixed with formaldehyde is very poor (fig. 1, E). By shadowing ofuntreated specimens a substructure is revealed (fig. 2, B): the membranes arecomposed of irregularly shaped units (linear dimensions about 20 to 40 m/x).

FIG. I (plate). A, section of membranes fixed with osmium tetroxide, showing a three-layered structure (a). A few particles of the size of ribosomes are attached to the surfaces (»•).

B, section of membranes fixed with potassium permanganate, showing two parallel rows ofsmall particles (sp).

C, section of membranes fixed with formaldehyde, showing poor preservation.D, untreated membranes. The electron dense patches (ed) in this figure and in the follow ing

ones indicate that several layers of membranes are lying on top of each other. A few particles(r) of the size of ribosomes are attached to the membranes. The larger particles (Ip) areprobably aggregations of the smaller ones.

E, membrane fixed with formaldehyde. The contrast is very poor. A few particles of thesize of ribosomes (r) are attached to the membranes, and many smaller particles are also seen.

F, membranes fixed with osmium tetroxide. A few particles of the size of ribosomes (r)are attached to membranes.

FIG. I

K. DEUTSCH, W. KRAUSE, and S. ROSENTHAL

FIG. 2

K. DEUTSCH, W. KRAUSE, and S. ROSENTHAL

Fixation of isolated mitochondrial membranes 321

Inspection of specimens fixed with osmium tetroxide, potassium perman-ganate, or formaldehyde and subsequently shadowed shows that a great numberof small particles are attached to their surfaces; their diameter is about 8 vnyu(fig. 2, c to E). In specimens fixed with formaldehyde or potassium per-manganate the particles are not so clearly separated as in those fixed withosmium tetroxide. Sections of specimens fixed with potassium permanganateshow two parallel rows of small electron-dense particles, of diameter about8 m/u. (fig. 1, B). The space between the particles has about the width of amitochondrial membrane, and about the same electron density as the em-bedding medium. One has thus the impression that fixation by potassiumpermanganate reveals the particles that are attached to the surface of themembrane, but not the membrane itself. Sections of membranes fixed withosmium tetroxide show the well-known three-layered structure (fig. 1, A).In sections of specimens fixed with formaldehyde the outlines of the mem-branes are not clearly visible (fig. 1, c). The thickness of the membranes, asestimated by a study of sectioned isolated mitochondria that had been fixedwith osmium tetroxide, is about the same as that of mitochondrial mem-branes in sections of whole cells.

A comparatively small number of larger particles (about the size of ribo-somes) are attached to untreated specimens and to those fixed with formal-dehyde or osmium tetroxide, and also to sectioned specimens fixed with osmium(fig. 1, A, D to F); but they are not found in specimens treated with potassiumpermanganate specimens (fig. 2, A).

DiscussionThe large particles attached to untreated specimens and to those fixed

with osmium tetroxide or formaldehyde (fig. 1, A, D to F) are of the same sizeas ribosomes. They probably are in fact ribosomes. Their preservation byosmium tetroxide but not by potassium permanganate (fig. 2, A) suggests this(Bradbury and Meek, i960). Thus they actually constitute a contamination.

Sections of isolated membranes fixed with osmium tetroxide show the samestructure as the same membranes in sections of cells (Watson and Siekevitz,1956), and their thickness is about the same (fig. 1, A). Formaldehydefixation and presumably also the subsequent treatment preparatory to section-ing has a very adverse effect on the preservation of isolated membranes (andalso on membranes in cells) so far as electron microscopical work is concerned(fig. 1, c, E). In classical histology formaldehyde is considered to be a goodfixative, and, as our previous investigations on the influence of fixatives on

FIG. 2 (plate), A, membranes fixed with potassium permanganate.B, untreated, shadowed membranes, showing a coarse, mulberry-like structure,c, membranes fixed with osmium tetroxide and shadowed. A great number of small

particles (sp) are attached to the membrane.D, membranes fixed with potassium permanganate and shadowed. The mulberry-like

structure is probably produced by the aggregation of small particles in groups.E, membranes fixed with formaldehyde and shadowed. Small particles (sp) are attached to

the membranes. The larger patches (Ip) are probably aggregations of small particles.

322 Deutsch, Krause, and Rosenthal

proteins have shown (Deutsch, Fischer, and Krause), it also increases thecontrast of proteins (as osmium tetroxide does). But the membranes alsocontain phospholipids, which are fixed by osmium tetroxide but not byformaldehyde (Mercer and Birbeck, 1961). This difference may explain whythe membranes are very poorly preserved after treatment with formalin. Theadverse effect of formaldehyde fixation, however, can be counteracted by'staining' (Magnan, 1961). Thus the 'stain', besides increasing the contrast,also acts as a complementary fixative. The appearance of isolated membranes,whether sectioned or not, is very different from the appearance of membranesin sections of whole cells, when potassium permanganate is used as fixative(fig. 1, B). It is very likely that the methods used for preparing isolated mem-branes bring about structural changes. In view of the investigation byBradbury and Meek (i960) it is not surprising that membranes treated withpotassium permanganate are not so well preserved as those treated withosmium tetroxide. According to these authors potassium permanganate isnot a true fixative.

The substructures of the membranes are revealed by the shadowing ofuntreated membranes (fig. 2, B). The units observed in untreated shadowedspecimens have about the same size as the electron transport particles ('ETP')(Green and Oda, 1961; Ziegler, Linnane, Green, Dass, and Ris, 1958). Theirshape is somewhat different, but this may be the result of difference inpreparation, if, indeed, the units observed by us are identical with the ETP.

The particles attached to the membranes, as observed in specimens fixedwith osmium tetroxide, potassium permanganate, or formaldehyde and sub-sequently shadowed (fig. 2, c to E), are of special interest. These particlesare also seen in specimens fixed with potassium permanganate and thensectioned (fig. 1, B). They have roughly the same size as the EPs (elementaryparticles) described by Fernandez-Moran (1962, 1963) and by Green, Blair,and Oda (1963), and it is quite possible that they are identical with them.Particles of about the same size attached to the mitochondrial membrane havealso been described by Stoekenius (1963). But we have not found anyevidence for the existence of the stalks, by which, according to Stoekenius,they are attached to the membranes. The resolution of the instrument thatwe have used is presumably not adequate to show the stalks. Stoekenius, whostudied mitochondrial membranes after negative staining, found about 3,000particles per /J,2; we counted about 4,000 particles per /A2, though, as Stoeken-ius points out, the figures may not be very significant. He claims that fixationdestroys these particles. It should, however, be pointed out that, so far as weknow, he only investigated unshadowed specimens, and with unshadowedspecimens our findings agree with his. But in view of the evidence providedby shadowed specimens and by those fixed with potassium permanganate andthen sectioned, we think it more likely that the contrast in unshadowedspecimens is not sufficient to reveal the particles.

Our results support the findings of Fernandez-Moran, Green, and Stoe-kenius, since we have also found that a great number of small particles are

Fixation of isolated mitochondrial membranes 323

attached to the surfaces of the membranes. We think it very likely that theparticles described by the authors named are essentially identical with ours.The proof, however, rests only on morphological evidence. We hope, byfurther investigations, to contribute to the understanding of the biologicalsignificance of these particles.

The authors wish to express their gratitude to Mr. E. Fischer, HeadTechnician of the Laboratory of Electron Microscopy, Greifswald University,who took the electron micrographs, and to Mrs. A. Krenz, who prepared thesections.

ReferencesBradbury, S., and Meek, G. A., i960. Quart. J. micr. Sci., 101, 241.Deutsch, K., 1957. Proc. Roy. Physical Soc. Edinburgh, 26, 7.

Fischer, E., and Krause, W., 1964. Quart. J. micr. Sci. (In press.)Egger, E., and Rapaport, S. M., 1963. Nature, zoo, 240.Fernandez-Moran, H., 1962. Circulation, 26, 1039.

1963. McGraw-Hill yearbook of science and technology, p. 323. New York (McGraw-Hill).

Green, D. E. and Oda, T., 1961. J. Biochem., 49, 742.Blair, P. V., and Oda, T., 1963. Science (Washington), 140 (Abstracts).

Lehninger, A. L., Wadkins, C. L., Cooper, C, Devlin, T. M., and Gamble, J. L., 1958.Science (Washington), 128, 450.and Wadkins, C. L., 1962. Ann. Rev. Biochem., 31, 47.

Magnan, C, 1961. Traite de microscopie e'lectronique, p. 912. Paris (Hermann).Mercer, E. H. and Birbeck, M. S., 1961. Electron microscopy. Oxford (Blackwell).Rosenthal, S., and Rapaport, S. M., 1964. (In press.)Schneider, W. C, and Hogeboom, G. H., 1951. Cancer Res., 11, 1.Stoekenius, W., 1963. J. Cell Biol. 17, 443.Watson, M. L.( and Siekevitz, P., 1956. J. biophys. biochem. Cytol., 2, 639.Ziegler, D. M., Linnane, A. W., Green, D. E., Dass, C. M. E., and Ris, H., 1958.

Biochim. biophys. Acta, 28, 524.